concurring in part and dissenting in part:
I. United States and Arizona Constitutional Contract Clause Claims
I concur with the majority’s conclusion that “Pure Wafer does not have a claim under the Contract Clause” of the United States or Arizona Constitutions. Maj. Op. 953. Thus, I agree that the judgment must be reversed and remanded.1
II. Arizona Breach of Contract Claims
However, I must dissent from the majority’s sua sponte decision to reach Pure Wafer’s alternative claims that the City breached the Development Agreement. Instead, we should remand for the district court (1) to consider whether to exercise supplemental jurisdiction over the breach of contract claims, and (2) (if it decides to exercise such jurisdiction) to make findings of fact as to those claims.
A. We must allow the district court to assess in the first instance tvhether to exercise its supplemental jurisdiction.
In holding the Contract Clause inapplicable to this case, we have dismissed the only federal claim before us. Thus, only claims for breach of contract under Arizona state law remain. Pure Wafer invoked supplemental jurisdiction over these claims in the district court but did not plead diversity. Therefore, the first question must be whether the federal, rather than the Arizona state, court should decide the remaining claims. See Carnegie-Mellon Univ. v. Cohill, 484 U.S. 343, 350 n.7, 108 S.Ct. 614, 98 L.Ed.2d 720 (1988). The majority errs in failing to allow the district court to consider in the first instance whether to exercise its supplemental jurisdiction.
Although district courts generally have supplemental jurisdiction over state law claims forming “part of the same case or controversy’ as federal claims, there are a number of circumstances in which “[t]he district court may decline to exercise [this] jurisdiction,” including when all federal claims have been dismissed. See 28 U.S.C. § 1367(a), (c). The decision is discretionary, and, if one of the § 1367(c) circumstances is present, “the exercise of discretion [is] triggered.” Exec. Software N. Am., Inc. v. U.S. Dish Court for Cent. Dist. of Cal., 24 F.3d 1545, 1557 (9th Cir. 1994), overruled on other grounds by Cal. Dep’t of Water Res. v. Powerex Corp., 533 F.3d 1087 (9th Cir. 2008). Section 1367 plainly vests this discretion with the district court. See 28 U.S.C. § 1367(c); see also Foster v. Wilson, 504 F.3d 1046, 1051 (9th Cir. 2007) (“The decision whether to continue to exercise supplemental jurisdiction over state law claims after all federal claims have been dismissed lies within the district court’s discretion.” (emphasis added)); Exec. Software, 24 F.3d at 1557 (“[S]ubsection (c) [of § 1367] requires the district court, in exercising its discretion, to undertake a case-specific analysis.” (emphasis added) (quoting H.R. No. 734, 101st Cong. § 29 (1990))); Imagineering, Inc. v. *960Kiewit Pac. Co., 976 F.2d 1303, 1309 (9th Cir. 1992) (providing that once all federal claims are dismissed, the exercise of jurisdiction over state law claims “is within the discretion of the federal district court” (emphasis added)), overruled on other grounds by Diaz v. Gates, 420 F.3d 897, 900 (9th Cir. 2005) (en banc).
Once it dismisses all federal claims before it, a federal court “must reassess its jurisdiction by engaging in a pragmatic and case-specific evaluation of the myriad of considerations that may bear on the determination of whether to exercise supplemental jurisdiction.” 16 James WM. Moore et al., Moore’s Federal Practice § 106.66[1] (3d ed. 2016). Such considerations include “economy, convenience, fairness, and comity.” Exec. Software, 24 F.3d at 1557 (quoting Imagineering, 976 F.2d at 1309). These factors must be “weighted] in each case, and at every stage of the litigation,” City of Chi v. Int’l Coll. of Surgeons, 522 U.S. 156,173,118 S.Ct. 523,139 L.Ed.2d 525 (1997) (quoting Carnegie-Mellon, 484 U.S. at 350, 108 S.Ct. 614), and the district court is in the best position to weigh them, see Hoeck v. City of Portland, 57 F.3d 781, 785-86 (9th Cir. 1995); 16 Moore, supra § 106.66[3][a]. In my view, an Arizona court would be better suited to adjudicate Pure Wafer’s claims based on Arizona law. However, the district court should make that decision.
Because the discretion lies with the district court and it is in the best position to make the decision, several of this court’s opinions indicate that we should allow the district court the first opportunity to consider the issue. See Watison v. Carter, 668 F.3d 1108, 1117 (9th Cir. 2012) (“On remand [after appellate ruling negated the district court’s basis for dismissing state law claims], the district court ... shall decide anew whether to exercise supplemental jurisdiction.”); Mendoza v. Zirkle Fruit Co., 301 F.3d 1163, 1174-75 (9th Cir. 2002) (“The decision to exercise [supplemental] jurisdiction remains discretionary with the district court.... We therefore remand for the district court to determine, in the first instance, whether the application of the Gibbs standard permits the exercise of supplemental jurisdiction, and to exercise discretion over whether such jurisdiction would be appropriate in the context of this litigation.”); Webster v. Omnitrition Int’l, Inc., 79 F.3d 776, 790 (9th Cir. 1996) (“The Attorney Defendants ask us to dismiss the state law claims against them for lack of pendent jurisdiction. The district court may, in its discretion, refuse to exercise supplemental jurisdiction after considering 28 U.S.C. § [1367].[2] We mil not examine the necessary factors in the first instance.” (emphasis added)).
In Fang v. United States, the plaintiff filed federal and state law claims against the United States based on her daughter’s death in a national park. 140 F.3d 1238, 1240 (9th Cir. 1998). The district court granted summary judgment on the federal claims for lack of subject matter jurisdiction under the Federal Tort Claims Act. Id. It also dismissed the state law claims for lack of subject matter jurisdiction under 28 U.S.C. § 1367(c)(3). Id. On appeal, after deciding to reverse the' district court’s dismissal of the federal claims, we considered the dismissal of the supplemental claims. Id. at 1241-43. We reasoned that, because the federal claims “were erroneously dismissed, the reason for dismissing the remaining supplemental claims no longer exist[ed].” Id. at 1244. Declining to rule on the defendants’ arguments that *961complex state law questions and predominance of state law issues called for us to uphold the dismissal of the state law claims, we held:
The decision to exercise supplemental jurisdiction is within the discretion of the district court and that court must be given an opportunity to make that decision. We therefore remand the case to the district court where it can determine whether it should retain jurisdiction over the state law claims in light of [these] alternate arguments.
Id.
In Hunsaker v. Contra Costa County, the plaintiff brought disparate impact disability claims under both federal and state law, seeking a permanent injunction. 149 F.3d 1041, 1042 (9th Cir. 1998). The district court ordered the injunction on the federal claim. Id. This court reversed, holding that there was no violation of the federal law. Id. at 1044. Over the plaintiff’s argument that we should, nonetheless, uphold the injunction under the alternative state law claim, we held, “[t]he district court did not rule on this claim, and we have nothing to review. We should allow the district court to consider this claim in the first instance or, in its discretion, decline to exercise supplemental jurisdiction.” Id.
Here, the district court initially exercised supplemental jurisdiction over the state law breach claims but ultimately dismissed them as moot, based on its ruling on the constitutional claims. As in Fang, our disposition of the federal claims on appeal negated the basis on which the district court dismissed the state law claims. In Fang, we could have ruled on whether the plaintiffs claims presented novel state law issues that substantially predominated over the federal claims, which could have disposed of the state claims. However — recognizing (1) that deciding whether to exercise supplemental jurisdiction is a discretionary question for the district court, and (2) ruling on the defendants’ alternative arguments would have deprived the district court the opportunity to decide in the first instance whether to exercise that jurisdiction — we remanded so the court in the best position to rule on the jurisdiction issue could do so. See Fang, 140 F.3d at 1244. The present case compels the same result. As in Hun-saker, the district court did not rule on the alternative state law basis for the judgment. 149 F.3d at 1044. Therefore, we have nothing to review with respect to that alternative basis, and “[w]e should allow the district court to consider [the breach of contract claims] in the first instance or, in its discretion, decline to exercise supplemental jurisdiction.” See id.
B. Because the district court held that the breach claims were moot, it made no factual findings specific to those claims, and we err in failing to remand for factual findings noio that the claims are no longer moot.
Even if we knew the district court would decide to exercise supplemental jurisdiction, remand would still be necessary, because the district court made no factual findings with respect to the breach of contract claims.
Although pure interpretation of language in a contract is a question of law, see Grosvenor Holdings, L.C. v. Figueroa, 222 Ariz. 588, 218 P.3d 1045, 1050 (Ct. App. 2009), the primary purpose of contract law in Arizona is to determine the parties’ intended meaning of the contract at the time of formation, see Taylor v. State Farm Mut. Auto. Ins. Co., 175 Ariz. 148, 854 P.2d 1134, 1138-40 (1993) (in banc), which “is a question of fact left to the fact finder,” Chopin v. Chopin, 224 Ariz. 425, 232 P.3d 99, 102 (Ct. App. 2010). In cases *962where extrinsic evidence is used to determine the parties’ intent from multiple reasonable interpretations of contractual language, the interpretation of the contract will generally also become a question of fact. See In re Estate of Pouser, 193 Ariz. 574, 975 P.2d 704, 709 (1999); Taylor, 854 P.2d at 1144-45. Likewise, “[w]hether a party has breached [the] contract is a question of fact.” Great W. Bank v. LJC Dev., LLC, 238 Ariz. 470, 362 P.3d 1037, 1045 (Ct. App. 2015). We must allow the district court an opportunity, in the first instance, to make these factual determinations for Pure Wafer’s breach of contract claims.
In its order, the district court found the breach claims moot in light of how it disposed of the constitutional claims. This treatment of the breach claims resulted in the court devoting only about one-quarter of a page (out of a thirty-two page opinion) directly to these alternative claims. And even that minimal space includes no findings of fact specific to the breach claims, instead providing that the district court “need not reach Pure Wafer’s alternative [contract] claims.” Because the district court found the breach claims moot based on the premise that the constitutional claims were valid, our ruling today has removed the only basis on which the district court dismissed those claims. Therefore, the breach claims must be revived. See Fang, 140 F.3d at 1243-44 (reinstating state law claims where, after court of appeals’ ruling, district court’s “reason for dismissing the remaining supplemental claims no longer exist[ed]”). Although the breach claims are revived — because (1) we have negated the only conclusion the district court reached with respect to these claims, and (2) the district court made no factual findings specific to the breach claims — there is nothing for this court to review concerning the claims at this stage. Instead, by deciding the breach claims, the majority steps into the role of fact-finder, see United Cal. Bank v. Prudential Ins. Co. of Am., 140 Ariz. 238, 681 P.2d 390, 454 (Ct. App. 1983) (resolving evidentiary questions is the role of the district court), without the benefit of district court findings concerning these claims, and relying only on findings the district court made in the context of constitutional claims that we hold today to be invalid.
C. The majority decides this case based on the breach of contract claims it believes Pure Wafer should have alleged, instead of addressing the claims Pure Wafer actually alleged.
In addition to assuming the district court’s role of factfinder, the majority’s breach of contract analysis fails to even mention the only two provisions Pure Wafer actually alleged in its complaint that the City breached. The complaint alleged the City breached the Agreement (1) under section 14.14 by failing to exempt Pure Wafer from the 2013 ordinance, and (2) under section 9.2 by failing to exempt Pure Wafer from the 2013 ordinance. According to the majority, Pure Wafer alleged “that the City, by enacting the Ordinance, had committed at least two different breaches of contract.” Maj. Op. 950. Although this initial statement about the breach claims is accurate, when the majority actually discusses the substance of those claims it does not once cite to, or analyze the content of, section 14.14 or section 9.2 of the Agreement. The majority also asserts that “Pure Wafer included a claim for simple breach of contract in its suit against the City, alleging that ‘Pure Wafer cannot comply with the Ordinance without incurring substantial costs which the Development Agreement allocated to the City.’” Id. at 952. While, again, the majority correctly recognizes that Pure Wafer included *963breach of contract claims, it implies, incorrectly, that Pure Wafer alleged facts concerning cost allocation specifically in the context of its breach claims. Pure Wafer did not. Instead of considering the breach claims as Pure Wafer alleged them, the majority analyzes the breach claims it believes Pure Wafer should have alleged, rules in Pure Wafer’s favor on those claims, and never addresses the breach claims Pure Wafer actually alleged in its complaint.
D. The majority’s analysis of the breach claims contradicts a plain reading of the Agreement, fails to address several relevant provisions, and demonstrates that further factual development may be needed.
As noted, the primary goal of Arizona contract law is to determine the parties’ intended meaning of their agreement and to give effect to that contractual intent. See Taylor, 854 P.2d at 1138-40. Thus, in attempting to discover the parties’ intent, courts are to liberally consider extrinsic evidence to show the parties’ intended interpretation of their contractual language. See id. at 1138-41. However, because “[i]n-terpretation is the process by which we determine the meaning of words,” see id. at 1138, extrinsic evidence is useful only to the extent it reveals how the parties’ intent is reflected in the words they chose to memorialize their agreement, see id. at 1140-44; see also Smith v. Melson, Inc., 135 Ariz. 119, 659 P.2d 1264, 1266 (1983) (in banc) (“A contract should be read in light of the parties’ intentions as reflected by their language and in view of all the circumstances.”). And the further an interpretation gets from the contract’s actual language, the more convincing the extrinsic evidence must be to show the parties intended that meaning. See Taylor, 854 P.2d at 1139^40. In addition, the parties’ intent for specific contractual language must be determined in light of their entire agreement. See Smith, 659 P.2d at 1267.
At the heart of the parties’ dispute are two difficult, interrelated issues: (1) the limits of Pure Wafer’s right to discharge into the City’s sewer, and (2) who bears the cost of bringing that discharge into compliance with Arizona Department of Environmental Quality (ADEQ) requirements. The majority appears to rely exclusively on section 4.2 of the Agreement to answer these questions. Although its interpretation could be valid, it fails to analyze the specific language of the provision on which it relies. Further, the majority fails to consider several other provisions implicated here. When read as a whole, the Agreement is clearly susceptible to multiple reasonable interpretations with respect to these issues.
1. The majority’s interpretation of Pure Wafer’s discharge rights contradicts section Jf.2’s plain meaning and has insufficient supporting extrinsic evidence to overcome that contradiction.
The parties dispute what limits the Agreement places on Pure Wafer’s right to discharge, and the City’s obligation to accept, wastewater into the City’s sewer system. Section 4.2 requires the City to provide Pure Wafer 195,000 gallons of daily “sewer capacity.” It further provides that the “City shall not reclassify [Pure Wafer’s] effluent” for purposes of sewer usage rates “unless there is a material change in the waste water quality from the specifications attached hereto as Exhibit F.” Exhibit F, in turn, lists 50 mg/L as the “typical value” for fluoride and 100 mg/L as the “maximum.”
Contrary to the findings of the district court, the majority concludes Pure Wafer only has a right to discharge wastewater of up to 100 mg/L fluoride. See Maj. Op. 947, *964956-57, 957-58. The majority relies on Exhibit F for this limit. See id. at 947-48. It also relies on certain extrinsic evidence: (1) testimony that, at the time of the Agreement, Pure Wafer expected to discharge up to 100 mg/L fluoride and this understanding was critical to its negotiations, id. at 956-57; (2) a 2004 letter indicating that a City employee understood Pure Wafer to have such a right, id. at 957; and (3) letters from ADEQ to the City prior to the Agreement reflecting that the City knew it must obtain an aquifer protection permit, see id. at 957.
The district court below persuasively analyzed, and found antithetical to the Agreement’s plain meaning, the conclusion the majority now reaches as to Pure Wafer’s discharge rights. Though section 4.2 obligates the City to supply sewer capacity, it does not provide that Pure Wafer has a right to discharge wastewater (through that capacity) with any specified level of pollutants, or that there are any limits (regardless of fluoride content) on what Pure Wafer may send through its 195,000 daily gallons of sewer capacity.
Exhibit F’s plain language cannot be read to impose discharge limits, as the majority asserts. The district court persuasively rejected the argument that Exhibit F established any measure of Pure Wafer’s right to discharge effluent of a particular fluoride content, and instead, held that no contract provision could reasonably be interpreted to set such a limit. The Agreement refers to Exhibit F only once. A plain reading of that reference shows Exhibit F does not create any right to discharge a certain amount of fluoride, but rather, relates only to “sewer usage fees” Pure Wafer must pay for “sewer capacity” under section 4.2. This reading was confirmed by Pure Wafer’s only trial witness, who was involved in the original planning of the Prescott facility. He testified that Exhibit F did not relate to any right for Pure Wafer to discharge certain contaminant levels but, instead, related only to pricing.
Although the majority points to extrinsic evidence to support its position, it fails to address the interpretation problem pointed out by the district court — that there is no provision in the Agreement, the words of which we could reasonably interpret to impose a 100 mg/L limit. There is no better indication of contractual intent than a plain reading of the language the parties chose to express their rights and obligations in the Agreement. And, here, that language does not support the majority’s position. Extrinsic evidence is persuasive only to the extent that it shows the parties’ intent through the meaning of the contractual language, see Taylor, 854 P.2d at 1140-44; Smith, 659 P.2d at 1266, and the further an interpretation varies from the written language, the more convincing the evidence must be to show the parties intended the proffered interpretation, see Taylor, 854 P.2d at 1139-40. The evidence on which the majority relies does not clear this bar. Neither does the majority address the testimony of Pure Wafer’s only witness, which contradicts the majority’s reading of Exhibit F. Based on the conflicting interpretations of the parties’ intended rights and obligations concerning discharge, at the very least, the district court should be permitted to determine the facts in the context of Pure Wafer’s breach of contract claims. And, if on remand, there is not persuasive evidence as to the parties’ actual intent, any remaining ambiguity in the parties’ rights and obligations should be interpreted against Pure Wafer as the party who drafted the agreement. See Polk v. Koerner, 111 Ariz. 493, 533 P.2d 660, 662 (1975). Interpretation against the drafter is particularly applicable where the “party is attempting to impose an obligation on another where otherwise such an *965obligation would not exist.” United Cal. Bank, 681 P.2d at 412.
2. The majority’s interpretation of the Agreement’s cost allocations contradicts the plain meaning of section ⅛.2 and ignores several provisions relevant to the financial obligations.
The parties’ discharge rights and obligations also relate to a second question at issue here: When environmental laws allow the City to release wastewater into the aquifer only if its fluoride content is below a certain level, and it is undisputed that Pure Wafer’s discharge causes excess fluoride levels, who bears the cost of abatement? The majority interprets section 4.2 to impose the cost on the City. See Maj. Op. 946-47, 954-57. It posits that, by agreeing to take Pure Wafer’s discharge, the City agreed to bear the cost if the law required purification to reduce the fluoride. However, this reading contradicts section 4.2’s plain meaning and fails to consider several other provisions that appear to affect the parties’ financial obligations.
Section 4.2 expresses when and how (logistically) the City provides “sewer capacity” to Pure Wafer: It must hold 195,000 gallons of physical carrying capacity in “reserve in its sewer disposal system at all times after commencement of construction of the Facility.” Section 4.2 refers to existing “[tjrunk line facilities ... currently in place [that] appealed] adequate” to provide this carrying capacity. However, if the trunk line facilities “prove[d] inadequate,” the “City [was] obligated to augment such facilities ... by constructing at no cost to [Pure Wafer] all mains, lines, and other facilities necessary to accept or accommodate the additional sewer flow or effluent from the facility.”
The majority suggests the City’s obligation to “augment [inadequate] facilities” amounts to a “regulatory contract” under which the City guaranteed Pure Wafer the benefit of the regulatory scheme existing at the time of the Agreement by agreeing to bear the cost of any changes. Maj. Op. 946-47, 954-57. However, a plain reading of section 4.2 requires only that the City make available a specified amount of physical space in the sewer system to accommodate Pure Wafer’s discharge. And, if the sewer lines existing at the time of the Agreement turned out to be insufficient to handle the required volume of discharge, section 4.2 placed the financial burden on the City to augment only the physical capacity of those lines to accommodate the additional discharge that could not otherwise physically fit through the system. See United Cal. Bank, 681 P.2d at 425 (explaining that where a contract includes general terms that accompany specific terms covering the same subject matter, “the meaning of the general terms is presumed to be limited [by] the enumerated specific terms and to include only those things of the same nature as those specifically enumerated”).
The majority again fails to explain how the language of section 4.2 supports its position. The section does not impose an obligation to build some extensive purification facility, unrelated to the sewer’s physical capacity to accept a certain volume of wastewater through its pipes. Some extrinsic evidence may support the majority’s position, but again, that evidence is persuasive only to the extent it can be tied back into the actual language of the Agreement, see Taylor, 854 P.2d at 1140-44; Smith, 659 P.2d at 1266, which the majority has failed to do. Instead, a plain reading of section 4.2 supports a more reasonable alternative. Given the conflicting interpretations, the district court should be given the opportunity to make findings of fact— in the context of Pure Wafer’s breach of *966contract claims — as to the parties’ understandings of the language in section 4.2. And if, at that point, ambiguity remains as to their intended meaning, that ambiguity should be resolved against Pure Wafer as the drafting party. See Polk, 533 P.2d at 662; United Cal. Bank, 681 P.2d at 412.
In addition to section 4.2’s plain meaning, several other provisions may bear on the parties’ financial obligations under the contract. Yet the majority has limited its analysis to section 4.2 in isolation. See Smith, 659 P.2d at 1267. Read as a whole, article 4 of the Agreement may also speak to the parties’ cost allocations. For example, the detail with which they allocate risk in section 4.3 tends to negate an interpretation that section 4.2 makes allocations not explicitly stated. Under circumstances not present here, section 4.3 expressly places on Pure Wafer:
responsibility] for any engineering and construction associated with the connection to that infrastructure[, which shall] include, but not be limited to, metering and sampling devices and structures, pipeline, pump stations, etc. ... [Pure Wafer] shall be responsible for sampling and testing costs.... In the event that [Pure Wafer] discharges effluent of an inferior quality than is required by permit, and the City’s facilities are negatively impacted, [Pure Wafer] shall be financially responsible.
Because some sections allocate the responsibilities, risks, and costs in such detail, it is reasonable to assume that the parties would have made similar explicit allocations in section 4.2, if they intended section 4.2 to have that effect.
The Agreement may also allow “surcharges” to Pure Wafer if its fluoride levels increase the City’s processing costs. Section 4.1, “Operations Water Supply,” provides that “Water Supply shall be at City’s sole expense, without any special assessments, costs, [or] surcharges.” Section 4.2 contains no protection against surcharges. Further, Pure Wafer’s only witness testified that section 4.2 and Exhibit F allow the City to charge higher sewer rates as discharge contaminants increase “because it will cost [the City] more to process.” His testimony indicates the parties may have understood, upon executing the Agreement, that if the quality of Pure Wafer’s discharge increased the City’s processing costs, the increased costs could be passed on to Pure Wafer.
While section 4.2 provides cost protections to Pure Wafer, these protections may not prohibit increased processing costs. Section 4.2 prohibits increasing Pure Wafer’s “sewer usage fees” absent “a material change in the waste water quality.” It also provides, “Sewer Capacity shall be at no cost” to Pure Wafer other than “normal sewer usage fees.” As noted, section 4.2 indicates “sewer capacity” refers only to physical sewer space to carry away discharge. Therefore, it would not preclude discharge-related charges for something other than that physical space, such as a surcharge for the City’s extra costs of processing the excess fluoride. Such a “surcharge” would not amount to a “normal sewer usage fee” protected under section 4.2, nor would it be assessed for “sewer capacity.”
In addition, evidence of the parties’ dealings as to the excess-fluoride costs may show them understandings of cost allocations. See United Cal. Bank, 681 P.2d at 418 (explaining that the parties’ treatment of terms after the contract is executed but before a dispute as to meaning arises “is entitled to great weight” as evidence of the parties’ intended meaning for those terms). During the period of their Agreement, the City accepted Pure Wafer’s discharge regardless of quality. However, whenever ADEQ found the City in violation of flúor-*967ide limits, the City demanded Pure Wafer’s help to remedy the problem. And the record reflects Pure Wafer did help. Although some evidence suggests Pure Wafer provided the assistance despite having no obligation to do so, other evidence suggests Pure Wafer believed it was obligated to help bear these costs: A 2004 letter from Pure Wafer to the City (following the City’s Notice of Violation (NOV) from ADEQ) outlined, “It is [Pure Wafer’s] intent to work with the City to assure that its discharge ... will enable the City to consistently meet all of its permit requirements, including fluoride.” The letter also explained Pure Wafer’s plan to reduce fluoride concentrations. Finally, the letter provided that “the NOV issued to the City is directly related to our operations and potential remedial solutions may require expenditures on our part.” Other 2004 letters indicate that Pure Wafer hired an environmental engineer to determine how to reduce fluoride levels. Pure Wafer’s reduction of fluoride when its discharge caused the City to exceed permitted levels tends to contradict Pure Wafer’s position that it had a right to discharge wastewater with any fluoride content. Given these interpretation problems, the district court should have the first chance to determine the facts as to all of the matters that may have affected the parties’ understandings of their cost allocations.
A final contract provision that may bear on the parties’ financial obligations is section 12, “Force Majeure.” Both parties claim it would cost several million dollars to achieve compliant fluoride levels. Given this burden, both parties might claim defenses under section 12, which protects a party from defaulting where “inability to perform [is] due [to] ... acts or the failure to act, of any utility, public or governmental agent or entity ... beyond the control or without the fault of such party.” Changes to environmental laws imposing new burdens on either party could possibly amount to governmental action beyond the control and without the fault of the parties. But the district court should be given the first opportunity to make factual determinations as to the parties’ understandings of the scope of this provision.
In sum, the majority errs in failing to consider (in any manner) many of the provisions that may bear on the issues before us and in failing to recognize that the provisions on which it relies are subject to competing interpretations by Pure Wafer and the City. We should allow the district court to make the factual determinations with respect to these issues in the first instance.
E. Without allowing the district court an opportunity to consider the question, the majority bases its decision on a record that may be insufficient to show the parties’ contractual intent.
In addition to the lack of written findings as to the breach claims, a review of the present record suggests that the abbreviated proceedings in the district court may not have allowed for admission of extrinsic evidence sufficient to show the parties’ contractual intent. The district court should be permitted on remand to determine whether, under Arizona contract law, the record is adequately developed with facts that show what the parties truly intended their agreement to mean.
1. Lack of attention to the breach claims indicates that the parties have not adequately developed the record as to those claims.
The history of this case evidences that the proceedings may not have sufficiently developed the record on the breach claims to allow a proper ruling on the merits. The *968parties conveyed their initial expectations for the timing of all the claims in this case in their joint December 2013 planning report, where they estimated that discovery would take until the end of April 2014 and the case would be ready for a three- or four-day trial at the end of May. Instead of progressing pursuant to this expected timeline, the evidentiary proceedings in the district court were completed within five weeks of the parties filing their initial planning report and concentrated almost exclusively on the constitutional claims.
Pure Wafer did not include breach of contract claims in its initial complaint. Although Pure Wafer later amended to add the breach claims at issue, it then immediately moved for a preliminary injunction, making no argument that its breach claims justified an injunction or that the claims would eventually succeed on the merits. Instead, its motion sought an injunction “pending a final judgment determining Pure Wafer’s constitutional claims” The motion referred to the breach claims only once, representing that Pure Wafer “add[ed] purely alternative claims for breach of contract ... and breach of the implied covenant of good faith and fair dealing ... [and that gjiven the City’s stated positions, those claims would not succeed and [were] asserted only in the alternative ... out of an abundance of caution.” It was with this perspective toward the breach claims that the parties and district court would have prepared for the preliminary injunction hearing, which the court scheduled for December 19, 2013.
The district court limited each party to only three and one-half hours to present evidence and argument. Thus, Pure Wafer called only one witness during its presentation-in-chief. At the close of the first day (during which the City also called only one witness), the district court suggested they proceed immediately with a trial on the merits, consolidate it with evidence heard on the motion for preliminary injunction, and finish presenting the evidence on the next available court date (January 14, 2014). The parties agreed to the district court’s suggestion. The court’s subsequent order provided that the remainder of trial would be limited to the City calling one more defense witness and Pure Wafer calling one rebuttal witness (the same witness it called in its direct presentation). The order also provided in a footnote that “the trial necessarily includes the issue of liability on any claims pleaded in the alternative, namely Pure Wafer’s alternative claims for breach of contract and breach of the implied covenant of good faith.” In other words, trial of the alternative claims was ordered only after Pure Wafer completed its presentation-in-chief, which focused primarily on the preliminary injunction on the constitutional issues. Indeed, when both parties would have been planning which witnesses to subpoena, which to call, what questions to ask, and what other evidence to present, they would have made those decisions understanding that their presentations needed to pertain only to Pure Wafer’s entitlement to a preliminary injunction based exclusively on its constitutional claims. And they certainly would not have made these decisions with the understanding that they needed to try their entire cases, which they had estimated (in their initial planning report) would not be ready for five more months. This approach to the evidentiary proceedings below highlights why the parties and district court did not attempt to develop the record for the breach of contract claims.
Like the evidentiary portion of trial and every prior filing in the case, the post-trial briefing and proposed findings and conclusions only superficially addressed the breach claims. As noted, the district court’s opinion devoted less than one-quarter of a page (out of thirty-two pages) to *969the breach claims, holding that they need not be reached and dismissing them as moot.
The parties’ treatment of the breach of contract claims on appeal farther supports the conclusion that the parties and district court did not attempt to develop the record for the purpose of addressing those claims. The only reference the parties’ briefing made to the claims was in the City’s opening brief: “Pure Wafer’s breach of contract claims ... were dismissed as moot and are not at issue on this appeal.” Pure Wafer never disputed this position. Indeed, when asked at oral argument, neither party was prepared to address the appeal as a breach of contract case. We asked the City, “is there a reason we can’t treat [this case] as a contract issue?” The City answered, “I don’t think the record is developed to that — that the findings of fact and conclusions of law are developed to that level to permit you to do that.”3
Pure Wafer was equally reluctant to agree that the case could appropriately be decided as a breach of contract claim on the present record. Even in response to the question, “can you still prevail in this particular appeal by persuading us that the district court made enough findings to establish that this was a breach of contract,” Pure Wafer never adopted that position. As its argument came to a close, Pure Wafer conceded that the parties’ cost allocations for reducing fluoride was a term fully covered by the Agreement. Following this concession, however, we clarified whether Pure Wafer thought that (because the term was covered by the Agreement) the case could be decided on a breach of contract theory, rather than by reaching the constitutional issue. Pure Wafer reluctantly answered, “Maybe.” This ambivalent response (the last word of Pure Wafer’s argument) typifies the parties’ and district court’s perspective of the breach claims throughout this case.
2. Having dearly not focused, on the breach of contract claims, the proceedings up to this point may have left the factual record without adequate evidence of the parties’ contractual intent.
In light of Arizona’s goal of giving effect to the parties’ intent, the Arizona Supreme Court in Taylor explained the great extent to which the state allows parties to offer extrinsic evidence to show their understandings of contractual language, and thus, their contractual intent. See 854 P.2d at 1138-41. In analyzing the meaning of language and how it shows the parties’ intent, Arizona has expressly rejected the requirement “to make a preliminary finding of ambiguity” before the court can consider extrinsic evidence. See id. at 1138. Instead, the court must consider all extrinsic evidence that may support a party’s reasonable interpretation of contractual language as showing the party’s theory of contractual intent. See id. at 1139. Under this more liberal approach, Arizona allows the court to “consider surrounding circumstances, including negotiation, prior understandings, ... subsequent conduct,” and the like in interpreting the contract. See id. at 1139-40; see also Smith, 659 P.2d at 1267 (‘When interpreting an agreement, the court may always consider the surrounding circumstances.” (citing Restate*970ment (Second) of Contracts § 212 (1981))). If extrinsic evidence shows the parties used language in their contract they mutually understood and intended to have a certain meaning, the court must give effect to that intent, even if the words have a different meaning under ordinary usage. See Taylor, 854 P.2d at 1139. “[T]he purpose is to produce the contract result the parties intended, not that which the judge intends. Some words are clear beyond dispute. Some may mean one thing to the judge but could have meant something else to the parties. It is the latter meaning that is important.” Id. at 1141 n.2.
Because the proceedings in the district court did not concentrate on the breach of contract claims, the parties had no reason to develop the record with all the facts relevant to those claims. Pure Wafer has consistently maintained that the parties extensively negotiated before executing the Agreement. It alleged that they met “on many occasions and communicated ... by telephone and written communications.” It also alleged they “spent several months negotiating the terms of [their Agreement] to provide the protection desired both to Pure Wafer and to the City, on the key economic development elements of their deal.” According to Pure Wafer, “effluent capacity and quality were material terms the parties considered and negotiated ..., and [the Agreement] reflects those bargains.” It alleged the parties negotiated for the City to accept effluent of a specific “chemical profile,” and to accept the risk that the City’s permit requirements could change, resulting in extra expense.
In describing the parties’ negotiations, the district court cited only one exhibit and seven pages of trial transcript, coming from the testimony of only one witness. And even these two items were considered by the district court only for their relevance to the constitutional claims. The court did not address how this evidence may have pertained to the breach of contract allegations, because it found those claims to be moot. Arizona law allows a party to offer — and requires the court to consider — all extrinsic evidence that supports the party’s reasonable interpretation of a contract. Given Pure Wafer’s continued insistence that the parties extensively negotiated the terms at issue in this case prior to executing the Agreement, it is most likely that more extensive and compelling evidence exists than the current record shows.
For example, there is no evidence in the record to establish the City’s intended meaning of the Agreement and understandings from the negotiations at the time of contract formation. Surely someone from the City involved in (or knowledgeable on the details of) the many months of “telephone and written communications” between the parties would be available to provide this information. Concerning negotiations, there may be several witnesses who can testify, early drafts of the Agreement, the parties’ notes on those drafts, communications regarding their understandings of certain terms, compromises they reached, terms they changed, etc., all of which is highly relevant and necessary to determining the parties’ intent, but none of which can be found in the record. The record indicates such evidence exists; it simply has not been made part of the record. For example, a 1997 internal Pure Wafer memo refers to discussions of sewer usage rates the parties had at “the Development Agreement meeting.” However, the record contains no minutes, notes, or correspondence to show specifically what was discussed at this meeting. Similarly, a 2011 letter from Pure Wafer to the City outlined that, in preparing the letter, Pure Wafer’s attorney “pulled [Pure Wafer’s] file from the time the Agreement was drafted and reviewed the notes of [his] *971meetings and telephone conversations with the City, as well as correspondence with the City and drafts of the Agreement.” Yet these notes, written correspondence, and draft Agreements are absent from the record. Given this record, the testimony of a single witness (testifying on Pure Wafer’s behalf) does not seem sufficient to establish the contractual intent of both parties at the time they executed the Agreement, especially where so much additional extrinsic evidence is likely available.
The parties’ true intent cannot be revealed without consideration of all available evidence and application of that evidence to the language of the contract as a whole. Though it is certainly within the fact-finder’s discretion, I anticipate the district court on remand would want to conduct additional proceedings with respect to the breach claims, because the lack of attention to those claims leaves the record far short of containing all available evidence of contractual intent. The parties should be permitted to develop concrete evidence of their positions and understandings of the relevant terms at the time they executed the Agreement.
III. Conclusion
The district court should have the first opportunity to consider whether to exercise supplemental jurisdiction, because we have dismissed the only federal claim. This dismissal also necessitates a remand for the district court to make findings of fact and conclusions of law specific to the breach of contract claims, having previously found those claims to be moot. The majority errs by stepping into the role of fact-finder in the first instance with respect to the breach claims; by failing to address the claims Pure Wafer actually alleged; and by failing to recognize that its analysis of the breach claims contradicts the Agreement’s plain meaning, fails to consider the Agreement as a whole, and is lacking adequate support from the record.
The circumstances warrant remand to permit the district court (or an Arizona court) the first opportunity to address the merits. Doing so would allow the trial court to focus on the language of the Agreement and, in accordance with Arizona law, consider any extrinsic evidence that supports a reasonably susceptible interpretation of the contract. While I am cognizant of the desire for a speedy and efficient resolution of this dispute, the majority’s opinion sacrifices a proper and thorough resolution of this case for a speedy one.
. I also agree that the City should continue to be enjoined from enforcing the Ordinance against Pure Wafer during subsequent stages of this litigation.
. Webster cites 28 U.S.C. § 1366, concerning "laws applicable exclusively to the District of Columbia,” which were not at issue in that case. The citation to § 1366 is clearly a typographical error that was intended to cite 28 U.S.C. § 1367.
. This questioning occurred during the City’s rebuttal. The panel initially raised the issue in its first question during Pure Wafer’s argument, only after the City had already completed its argument-in-chief without being asked to argue its position on the issue. Because the panel spent the majority of Pure Wafer’s argument exploring whether we could approach the case as a breach of contract dispute, it is troubling that the panel did not raise the issue at a time that would ensure the City an equal opportunity to address it.